Evaluation of fixed continuous positive airway pressure (CPAP) with C-Flex+ against fixed CPAP

In order to test the primary hypothesis the following PSG variables will be assessed: 1. AHI (total, REM and NREM) - total, obstructive, central, mixed and hypopnoea

Comparisons will be made using the above data. These continuous variables will be summarised using means (and standard deviations) or median (interquartile range), depending on the distribution. The within-patient differences between the modes will be presented for each of these variables using mean differences and 95% confidence intervals. The 95% confidence bounds (one sided) will be scrutinised to determine whether or not non-inferiority of the fixed CPAP with C-Flex+ can be declared.

In order to test the secondary hypotheses the following variables will be assessed:
1. SpO2 - Nocturnal oxygenation
1.1. Total time spent less than 90%
1.2. Lowest SpO2 during the night
1.3. Average SpO2 during the night
2. TST - Total sleep time
3. SE% Sleep efficiency
4. Sleep Architecture:
4.1. Min/% Non-REM sleep
4.1.1. Min/% N1
4.1.2. Min/% N2
4.1.3. Min/% N3
4.2. Min/% R sleep
4.3. Min/% Wake After Sleep Onset (WASO)
4.4. Arousals:
4.4.1. # of arousals/awakenings (all cause)
4.4.2. Arousals due to PLMS
4.4.3. Arousal Index (AI)
4.4.4. Arousals due to Respiratory Disturbance (RDI)
Comparisons will be made using the above data. These continuous variables will be summarised using means (and standard deviations) or median (interquartile range), depending on the distribution. The within-patient differences between the modes will be presented for each of these variables using mean differences and 95% confidence intervals. The 95% confidence bounds (one sided) will be scrutinised to determine whether or not non-inferiority of fixed CPAP with C-Flex+ can be declared.

5. Average Pressure Outputs will be summarised using means (and standard deviations) or median (interquartile range), depending on the distribution. The pairwise within-patient differences between the modes will be presented as mean differences and 95% confidence intervals, and the single sample t-test will be used to assess whether the differences between the modes are statistically significant.
6. Comfort - The 10cm comfort visual analogue scale will be summarised for each of the modes from data collected on nights 1 and 2 of the study using means (and standard deviations) or median (interquartile range), depending on the distribution of the data. The pairwise within-patient differences between the modes will be presented as mean differences and 95% confidence intervals, and the single sample t-test will be used to assess whether the differences between the modes are statistically significant.
7. Preference - The number (and percentage) of patients who prefer each of the modes will be presented, along with 95% confidence intervals for these percentages and McNemar's test using a X2 test in a 2 by 2 contingency table will be used to assess whether the differences between the modes are statistically significant.
8. Breathing event output - Device reported scoring will be compared to both standard and specialised PSG scoring. The correlation coefficients will be determined. The PSG's from the study night will be considered the gold standard for identifying and quantifying apnoeas and hypopnoeas during sleep. The accuracy of the CPAP device in detecting residual events (clear airway apnoea, obstructed airway apnoea, hypopnoea, apnoea hypopnoea index, respiratory effort related arousals and Cheyne Stokes Respiration) will be based on comparisons of the CPAP and the PSG by correlation and agreement using the method of Bland and Altman. During these analyses, a clear airway apnoea will be considered a surrogate of a central apnoea and an obstructed airway apnoea a surrogate of an obstructive apnoea. Mixed apnoeas will be classified as central on the PSG recording. For the AHI, Receiver-Operator Characteristic (ROC) curves will be constructed to determine optimal cut-off values for determining positive and negative likelihood ratios. A log transformation of the data may be performed to improve the scatter of the differences as the AHI increases. Patients with and without residual OSA on CPAP will be compared using a two sample t-test.
9. Other relevant statistical tests including multivariate analysis may be performed according to the discretion of the Principal Investigator and Statistician.

Any deviations from points 1 - 8 in this plan will be reported in the final manuscript.